Purpose Despite stage migration as a result of screening, many individuals are diagnosed each year with metastatic (M+), as opposed to localized (M0), prostate cancer. This study describes features that characterize patients with M+ compared to those diagnosed with M0 disease. Materials and Methods Patients enrolled in the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE™), a national, longitudinal registry of men with prostate cancer, formed the basis of this study. The prevalence, and changes with time, of patients with M+ and M0 cancer by clinical and sociodemographic characteristics were examined. Results Of 10,113 patients diagnosed between 1990 and 2003, 266 (2.6%) had M+ disease at diagnosis. From 1990 to 1997, 4.2% of 4020 total patients had M+ versus 1.6% of 6093 total patients diagnosed between 1998 and 2003 (odds ratio 0.34; 95% confidence interval 0.24−0.48; P < 0.0001). In univariate analysis, advanced age, higher prostate-specific antigen, Gleason grade, black race, lower income, and lower educational level were associated with M+ versus M0 disease ( P < 0.01). However, in multivariate analysis, only higher serum prostate-specific antigen and higher Gleason grade, and not the sociodemographic variables, remained associated with M+ disease ( P < 0.01). Patients with M+ diagnosed between 1998 and 2003 are more likely to harbor high-grade (Gleason ≥8) primary tumors (62% vs. 45%, P = 0.02) than those diagnosed between 1990 and1997. No changes in age, race, education, insurance status, or income were observed in the early versus late era. Conclusions These findings show a reduction in the incidence of metastatic disease at initial prostate cancer diagnosis. Furthermore, biologic, rather than socioeconomic, factors are associated with this type of disease presentation.